Open FlyingWombat opened 2 years ago
Looks similar to this: https://github.com/umlaeute/v4l2loopback/issues/476
Ya, IBT is the culprit: booting with ibt=off worked. Similar problem reported here too, maybe we can see how they deal with it: https://github.com/NVIDIA/open-gpu-kernel-modules/issues/256
Edit Kernel aborts mount with "Kernel BUG" traceback when IBT is enabled. Indirect Branch Tracking (IBT) was introduced in kernel 5.18 with 11+ gen Intel CPUs.
A workaround is to set ibt=off kernel parameter. Compiling with -fcf-protection=branch did not fix.
If this is an upstream issue, feel free to close.
Original text:
I moved my hard drives to a new PC and rebuilt the kernel with the exact same config as before. Got this kernel bug when I tried to mount. All 3 drives are readable with bcachefs show-super and dd, so they're not broken.
Dmesg output:
Version Linux v5.18.10 bcachefs commit 7a90ae891e4ad96fe2d665c88088730c98c4e79e bcachefs-tools commit bad0c8c50758b4447d529f61017c1a8c85976a3e
Generic info superblock info: bcachefs-general-info.txt
system info (that's changed): cpu: intel i7-12700k (alder lake) mb: MSI Pro z690-a wifi ddr4 boot drive: nvme-Samsung_SSD_970_EVO_Plus
Building kernel with bcachefs debug options yielded a slightly different call trace: bcachefs-5.18.10-g7a90ae891e4a-debug-trace.txt